Clone allways performed on last device created (with jbconfig)

There is no reason to delete libraries so many times

Cloning will depend on what is the source and what is the target set - it is simple as that.

You should use the clone storage node field and label the clone pool media on the storage node.

Reconfiguring jukebox again & again would not help.

Hello,

The clone storage node field, is not taken into account.

Do you agree that if I take a backup on the backup server, the clone must be on the backup server and not on a storage node ?

That's my case !

Regards,

Hello,

How can I force source and target to the backup server, cloning from storage node to backup server takes more time.

Thank you
Regards,

It is, but not for clients but rather storage nodes - please see documentation.


I don't, but that entirely depends on your setup. Please see documentation and see with your support if something is not clear (this is covered well in documentation).

Hello,

I read in the documentation:

Determining the source for reading the volume

The following logic is used to determine the read source for the cloning operation:

1. If the source volume is mounted, the storage node of the device on which the volume is mounted is used as the read source.

2. If the source volume is in a media library, then the Read Host attribute in the Library resource is used as the read source.

3. If the source volume is in a media library and the Read Host attribute is not set, the host of the first device in the library is used as the read source.

4. If the source volume is not mounted and is not in a media library, then the NetWorker server is used as the read source.

---

I should be in the case number 3, it always uses the last drive created in my case a storage node.

Regards,

Have you checked read hostname?

Hello,

Both clone storage node and read host name are empty.

Regards,

It must be empty, if not any restore will be through the "read hostname" (I had this surprise) even for restore performed by real storage nodes (for example Oracle database restore) then it goes through the network.

I've 2 drives.

Regards,

I never saw read hostname to be empty - are you sure? Anyway, how many devices do you have on server (I assume that is where you wish to make cloning)?

That's not my case, but you can easily fix that with restore storage node field if you want.

If you have two drives, one will be used for read and one for write - simply as that (read will be on node which has been used for write). I do cloning via script only and I always control this sort of things by setting read hostname for next nsrclone operation and I never had any issues - I would suggest same approach.

Hello,

I also use scripting, here's the command:

mminfo -q"savetime>=4/12/07,pool=MCBOFD,copies=1,!incomplete,valid,!suspect" -r"ssid" -omn >c:\batch\cloneout.mcb

nsrclone -v -S -b MCBCLD -f c:\batch\cloneout.mcb

Is there anything I could change ?
Regards,

Set read hostname via nsradmin.

Hello,

I've set it, it always asks for tape in the Networker Server, even for restore, which must be performed by storage nodes.

For the storage nodes, I've forced the "recover storage node" field to them self, it doesn't help.

I don't know what are the role for "clone storage node" and "recover storage node", they're not used.

Regards,

You should clone data from host from host on which data is saved to same host. Set clone storage node on storage nodes clients only to themselves. Have script function for each node and before each being execute modify read hostname accordingly. If that doesn't work your either there is a bug or something wrong with what or how you are doing it - in each case I would suggest to involve your support.